This application is submitted as one of two collaborating 4-year proposals under the Program Announcement "Collaborative Studies of Mental Disorders." The broad aims are to expand a unique, existing set of pedigrees and test candidate regions for linkage and association with bipolar disorder. For the Johns Hopkins site, this represents a revised competing renewal application while for the University of Chicago site this is a new proposal. Together, the two sites propose to double the existing family resource by ascertaining an additional 80 moderate-sized families through bipolar I probands with 2 or major siblings affected with recurrent major affective disorders. All participants will be interviewed by trained psychiatrists who have established excellent inter-rater reliability. Diagnoses will be assigned by 2 non-interviewing psychiatrists who review all clinical data. This sample has already proven to be of value to the field of psychiatric genetics. Among clinical findings that have spawned new research directions are the reports of anticipation, parent-of-origin effects, a high prevalence of BPII among the close relatives of bipolar I probands, and high rates of co-morbid panic disorder in a subset of families. The sample provided the first evidence of linkage to 18q, and the first molecular evidence for a parent-of-origin effect in bipolar disorder. Findings from prospective studies of the 2nd half of this sample demonstrate high allele sharing between bipolar II sib pairs at the several loci in 18q21. Exploratory analyses of co-morbid panic disorder and alcoholism have suggested methods for predicting heterogeneity between bipolar families. In addition to collecting more families, the investigators propose to genotype the existing and additional family sets at candidate regions implicated by a recent genome-wide scan for linkage that has been completed on 68 pedigrees from this sample. It is also proposed to use linkage and association methods to extract the maximal genetic information needed to locate genes influencing susceptibility to bipolar disorder. This is a carefully clinically assessed family set. The studies generated from this resource have demonstrated its value and have provided testable hypotheses for further work.